Storage is continuing to develop at a breathless pace-both in the use of advanced materials and in the way those materials are being applied.
New products and services involving NAND (Not AND) flash, Peripheral Component Interconnect express (PCIe) cards, image cloning, storage pooling, automation, improved capacity management and enhanced aerial density on digital tape are coming into the market from established companies and startups alike. These technologies will provide the foundation for future storage IT.
Even tried-and-true hard-disk drives and long-term table storage will remain in wide use for at least the next 10 to 20 years, as much as the solid-state storage makers would like to see both technologies fade away, according to most analysts. Enterprises have too much IT capital invested in those technologies, and they work well enough in their established applications. Besides, the newest storage technology isn’t ready for prime time yet.
Phase-change Memory in the Offing
One of the futuristic technologies is a hot, but not exactly new one called phase-change memory (PCM), a potential replacement for NAND flash solid-state storage. Around for more than 40 years, PCM is a key component of rewritable CDs, DVDs and Blu-ray storage disks that use laser optics. But only in the last decade has research by IBM, Intel, Micron (through its PCM-dedicated Numonyx arm) and Hynix showed that PCM also can serve well in the digital data storage sector.
Flash disks are limited to holding one bit of data per storage cell. About six months ago, however, IBM’s PCM research team in Zurich found a way to enable each PCM cell to hold multiple data bits securely; previously, bits often became lost or corrupt at unpredictable times.
“We’ve now addressed this problem, and we believe we’ve solved it,” Dr. Haris Pozidis, manager of memory and probe technologies at IBM Research in Zurich, told eWEEK.
Pozidis said this latest development can lead to solid-state chips that can store as much data as NAND flash disks, which now are up to 1TB in capacity, but deliver about 100 times faster data movement speed to go with a much longer lifespan. “Today’s enterprise flash can endure about 30,000 read/write cycles; today’s PCM chips can do in excess of 10 million cycles,” Pozidis said.
That’s magnitude-scale improvement. Performance like that looks awfully good to storage manufacturers and enterprise IT decision-makers. However, Pozidis and other experts have said that PCM is still three to five years away from being productized. There’s much more testing to do, and manufacturers have yet to figure out how to produce it on a mass scale.
In the future, we’ll see other developments, such as vastly improved scale-out file systems, longer-term disk archiving, better storage management control and greatly improved capacity on the media. Storage media, including spinning-disk hard drives, solid-state disks, digital tape and optical disks, continue to become more capacious (and run cooler)as engineers and manufacturers continue to achieve performance improvements. Markedly better networking and processing speeds are also instrumental in these upgrades.
Using today’s standard storage tech, there have been some advancements that bear notice.
NAND flash will continue its journey from handheld devices into the data center. “The use of NAND flash in general is a continuing enterprise trend,” said The 451 Group storage analyst Henry Baltazar. “MLC [multi-level cell] flash-most often used in consumer products-is coming more into use for enterprise applications. And it’s less expensive than SLC [single-level cell] flash.
“Take a vendor like SanDisk, for example. Their high-end SAS [serial-attached storage] drives using SLC are priced around $20 per gigabyte. If you look at the MLC variety, that’s priced around $10 per gigabyte.”
PCIe Moves to Mainstream
Also, PCIe cards are becoming mainstream. Analyst Jim Handy of Objective Analysis said his firm forecasts that the NAND flash-powered PCIe interface will become dominant in the enterprise solid-state disk market in 2012, with unit shipments greater than the combined shipments of its SAS and Fibre Channel counterparts.
In 2004, Intel launched PCIe, an expansion-card standard based on point-to-point serial links rather than a shared parallel bus architecture. It’s designed to replace the older PCI, PCI-X and AGP standards. PCIe-based flash storage has the ability to bypass traditional storage overhead by reducing latencies, increasing throughput and enabling efficient processing of massive quantities of data.
Virtual Machine Cloning
We’ll also be seeing something called “virtual machine image cloning” as an alternative to file-system and data-store snapshots. Oracle is ahead of the pack here. Its new VirtualBox virtualization package (v4.1 launched earlier this year) includes a new virtual machine cloning facility-one of the first on the market.
“Right now, when you have a virtual machine running, you create a snapshot, which is a child of the current virtual machine,” explained Wim Coekaerts, who serves as Oracle’s senior vice president of Linux and virtualization engineering. “But that’s not something that can independently grow afterward. With a clone, you have a new entity that can have its own life and, subsequently, its own snapshots.”
A snapshot is an object and a part of the virtual disk, so it can’t be copied onto any other servers and used in any way. Plus, users don’t have any visibility into it, Coekaerts said. Clones are completely new virtual disk objects, independent units that can have new lives of their own.
Automated disaster recovery-either on-premises or from a cloud service-is coming. In the past, reconnecting data stores with systems and getting those systems running after a power outage was done manually. However, software now available is smart enough to get large portions of a virtualized system back online much faster and with less effort. Dell EqualLogic, EMC Data Domain, Hewlett-Packard and VMware are some of the vendors that offer this.
Storage Pooling Gaining Momentum
Storage pooling is another hot topic. This approach to storage virtualization delineates specific areas of the storage system to be dedicated to specific data flows to enable more efficient multitenant service deployments, for example. Sepaton championed this early on, and other vendors are following suit.
Virtualized storage systems break files into chunks of data, which are dispersed into numerous data center or storage locations, and then reassemble them on demand. Keeping data file chunks closer together in pools is said to provide faster reassembly of file chunks.
All these advances in storage technology are giving IT managers many options when it comes to finding ways to improve data storage capacity and performance. It’s also giving them plenty to think about when they have to decide which option can provide the most efficient performance and is available today at an affordable price.
Editor’s note: This is the first of a three-part series. Sept. 26: What thought leaders are saying about the future of storage.